La Lega Toscana di Protezione

Learn more about:

The Project
The Team
The Methods
Our Resources

About our Approaches

Volume Transcription

The purpose for the transcription of the notebooks is to reproduce the original in a word document in order to facilitate its electronic ease of use. This involved both the reproduction of the general organization and structure of each page as well as preservation of all spelling (misspellings), punctuation, extraneous capitalizations, and any other characteristics of the page in order to render the word document as similar as possible to the source text. We followed certain transcription guidelines published online by the Metropolitan Museum of Art, the Smithsonian Institution, and various university library resources, which all stressed the importance of reproducing the original document as closely as possible in electronic form. We made extensive use of bracketed [tags] to mark parts of the text that were either illegible, crossed out, signed, underlined, or appeared in between or written over top of the other lines of text, in part to aid with the digitization of the texts.

Volume Translation

Roughly the first 75 pages of both the transcription and translation were completed by students taking an Italian Translation Workshop during the semester of Fall 2017. During Spring 2018 that transcription was gone over and corrected, and the existing translation, which had been a conglomerate of the different translations of various students, was revised and standardized in order to render it more uniformly. Continuing the work of last semester, the transcription and translation were completed through the end of the first volume of meeting minutes, which span from 1919 through 1925. Work on the second volume of minutes, spanning 1933 through 1940, is currently underway for potential study in the fall of 2018.

The use of a formal and repetitive language and tone in each meeting throughout the entire first notebook prompted the use of a similarly formulaic language within the English translation. We felt that the use of meeting-specific phraseology (e.g. the meeting was adjourned rather than the meeting is ended) was appropriate in this situation. Apart from these instances, the translation aimed at maintaining wording and general sentence structure except in places where doing so would threaten the clarity of the translation, in which case the English analogues of certain figurative phrases were employed.

Beyond the translation alone, the process involved the standardization of many commonly used words and phrases to ensure that they were translated in a codified manner throughout the entire volume of minutes. In order to maximize the comprehensibility of the translation, corrected spelling, capitalization, and punctuation were used for instances where they were incorrect or confusing in the source text. We also felt it would be helpful to have the lines in the translation correspond, if somewhat loosely at points, to those in the transcription, to help with navigation or comparison between the two documents, thus increasing their ease of use as a tandem resource as well as increasing their accessibility to those beyond the Italian-speaking community.

We found it especially helpful to create supplementary documents to help us keep track of certain things throughout this entire process. This includes a list of commonly used words and phrases, as previously mentioned, an index of names, which can be used to find instances where people were mentioned within the minutes, and an index of places encountered in the documents with their corresponding addresses when available.

Text digitization and encoding

The first of the two digital pillars of the Lega Toscana research project seeked to explore the first volume of minutes in its entirety and explore different interpersonal and role-based relationships between the members of the Tuscan League throughout the minute log. The expression of these relationships in the minutes was mostly gathered in a specific series of contexts. Originally, the motivation behind this exploring this specific document was to identify relationships and rapport established and reinforced by organizational members in the form of meeting proposals (i.e. who were the members who made motions during the meeting, and who were the people who supported them?), but as the encoding persisted, an interest in committee formations, dissemination of funds from the Lega, and activity of members across the organization began to surface. By the end of the semester, the research touched on each of those subjects, with the use of a handful of different analytical programs.

Text Encoding Initiative

The entirety of the Tuscan League's first volume of minutes were uploaded and analyzed in a software program called <oXygen/>, which is a textual analysis program that implements an extensible markup language to allow for encoded annotations or supplemental biographical information on Tuscan League participants for analysis. The markup is written to descriptively flag different objects of interest within the text, to allow for computational analysis at a later point.

The specific schema used for development on this project was written in compliance with the Text Encoding Initiative, which is an open-source collection of different elements, tagging conventions, and document types that are optimized for use in the humanities domain, and whose specialty is in manuscript digitization. One of the project mentors for this portion of research, Elisa Beshero-Bondar, is a member of the TEI Technical Council, the team tasked with maintenance and development of the TEI ecosystem. Elements were selected from the master list of TEI elements from the TEI All schema set, and customized for integration into the minute logs. Specific items that were encoded in the minutes included any list of member acceptances or departures, presiding officers, proposals of any kind, or grants of funds from the Tuscan League to individual members. Additionally, every iterance of an individual's name was tracked throughout the minutes as well, which assisted in the codification of individual names in spite of a lack of complete documentation on affiliates of the League. A sample portion of code below shows how proposals were tagged throughout the volume:

  <seg type="proposal" subtype="evento">Fu proposto da <persName ref="#zinim"
     role="proposer">Modesto Zini</persName> e assecondato da<lb/>
          <persName ref="#marconic" role="supporter">Carlo Marconi</persName> di dare
          un ballo ma non fu definita<lb/> la data</seg>
       <list type="committee" subtype="investigazione">
          <date when="1924-10"/>
          <note>Allora <rs ref="#pellegrinic">il presidente</rs> scielse il seguente
             comitato<lb/> per trovare il locale.</note>
          <item><persName ref="#pasquinellis">Santino Pasquinelli</persName><roleName
                type="committee" role="investigatore"/>.</item>
          <item><persName ref="#marconic">Carlo Marconi</persName><roleName
                type="committee" role="investigatore"/></item>
          <item>e <persName ref="#borellim">Martino Borelli</persName><roleName
                type="committee" role="investigatore"/>.</item>

In the code sample above, we show an example of how both proposals and lists are formulated. The participants in proposals were given key-value pairs that described the role they had in the proposal, whether it was he who made the motion, or he who seconded it. Additionally, the volume contains written proposals -- which were always proposed by at least three members -- and rejected proposals, which were processed differently than passed proposals (since they were not typically met with support by another member). The majority of information that was used to process the analyses of member activity and relationships in the Lega were mined primarily from different lists that appeared in the minutes and proposals. Lists were used to provide information on committee formations, officer elections and absences, member acceptances or departures, and sometimes even member compensation.

As noted earlier, the TEI All schema contains about 570 total elements, of which only 20 to 25 typically should appear in a DH project like this. As such, a customized ODD, or One Document Does it all file, was drafted to generate and enforce customized rules that applied specifically to this project. Additionally, another technology named Schematron was employed in the same document, which is another schema language that permits further constraint of possible values, patterns, and element hierarchies within the encoded document that are more difficult to express exclusively with an ODD file. Both the ODD file and the encoded transcription of the minutes are available in full at our GitHub repository, along with an introductory reading on our Wiki page.

Transformations with XSLT

When encoding texts with XML, there are other flavors of the XML language that can be used for a number of different purposes, from identity or text transformation to file conversion to visualization design. The central transformation language that was used in this project is XSLT, which can effectively process files a number of ways. XSLT is a file that receives an input file and performs transformations according to templates that are written into the program, and it can output the information into a host of different file types. In this specific research project, XSLT was used to transform the volume of minutes into other XML files that were used for the generation of a personography and data compilation; into HTML webpages that contained the Lega minute scans, transcriptions, and translations; and into graphic visualizations that help to narrate the different trends and traits expressed at the individual and group level in the organization. An example of some XSLT code is below.

      <xsl:template match="*" mode="flatten">
        <xsl:element name="{name()}">
            <xsl:attribute name="tagType" select="'startTag'"/>
            <xsl:attribute name="tagId" select="generate-id()"/>
            <xsl:copy-of select="@*"/>
                <xsl:when test="self::pb[ancestor::div[@type = 'transcription']]">
                    <xsl:attribute name="lang" select="'it'"/>
                <xsl:when test="self::pb[ancestor::div[@type = 'translation']]">
                    <xsl:attribute name="lang" select="'eng'"/>
        <xsl:apply-templates mode="flatten"/>
        <xsl:if test="not(self::lb | self::pb | self::date)">
            <xsl:element name="{name()}">
                <xsl:attribute name="tagType" select="'endTag'"/>
                <xsl:attribute name="tagId" select="generate-id()"/>
                <xsl:copy-of select="@*"/>

The code snippet above is an example from the XSLT file that tranformed the volume of minutes into an HTML reading view of the three different versions that we produced for this project -- the scanned images, transcription, and translation. This entire file, along with all other XSLT files, is availab on our GitHub page for study.

Social Network Analysis

In addition to the XML flavors that were used to analyze the volume of minutes quantitatively, the social network analysis program Cytoscape was used as a supplemental technology to generate social network graphs that plotted shared activity of members, specifically of those who were co-participants in a proposal during a meeting. The social networks generated for this research were primarily focused on identifying individuals who contributed or supported proposals during the meetings, and seeing if there were any notable trends or patterns among those members.

Social network analysis can potentially evoke three separate forms of netowrk importance, or centrality, with these computations: Degree centrality, betweenness centrality, and closeness centrality. Those three concepts, among other modes of network analysis, are explained in more depth in our Key Members webpage.

Spatial Mapping

The Tuscan League community was mapped using the software ArcMap and the experience afforded in pursuit of the Geographic Information Systems (GIS) certificate at the University of Pittsburgh. The primary objective of these maps is to highlight where Tuscan League community members lived and worked, and how that changed over time.

First, we assembled a database of about 400 street addresses, people and business names, and notes mentioned in the League's meeting minutes, applications to the Women's Auxiliary of the League, and our biographical research. We then created a geocoder which matched addresses in the database to spatial locations using street and address boundary data available on the Allegheny County GIS Database. The geocoder matched about 80% of the addresses. The remaining addresses had to be manually corrected for reasons like the spelling of the street had changed over time or the street address no longer existed. This was especially true for many addresses in the Hill District as it nears Downtown, where Wylie and Webster Avenues were truncated in the 1960s in order to build I-579. Those addresses were matched based on 1903 maps of Downtown Pittsburgh published on Historic Pittsburgh.

With all of the data points imported, we symbolized the residential locations by color and decade to see temporal patterns in residence location. A neighborhood layer was also added from the Allegheny County GIS Database to the map and labeled only neighborhoods within 0.2 miles of a residential or commercial point, to see which neighborhoods were most important to the community. Lastly, we analyzed the map using the ArcMap Spatial Analyst Point Density tool to visualize where residential points were most tightly clustered.

This map could be expanded upon by adding data from earlier time points that may show more of a migration from Downtown and the Hill District to the South Hills. There are historic reports of Italian - American communities Downtown and on the Hill, but location data is not yet available.

Lega Family Research

Biographies on the most influential members of the Tuscan League were drafted to highlight the lives of specific men, and how their lives were influenced by membership in the Tuscan League. Initially, an index created in the Fall of 2017 provided an inventory of members and their general relative importance. As men were mentioned in the volume, the page was noted in the member index. Some of the most active members throughout the volume included Casimiro Pellegrini, Muzio Frediani, Stefano Maffei, Emilio Marchetti, Michele Simonetti, Santino and Americo Pasquinelli, and Modesto Zini.Online databases like Historic Pittsburgh, Ancestry, Newspapers, and Family Search provided articles and more personal details through cencuses, draft registrations, death certificates, and biographical profiles.

The evidence compiled through this initial research helped to inform outlines of biographical information on the members. Some people, like Emilio Marchetti and Michele Simonetti, were very active in the Lega Toscana but not in the Pittsburgh community, according to the primary sources. The biographical summaries of each of the League's important families is the end-result of this more individually-oriented biographical research. Including information like immigration dates and occupations held, the biographies also touched on the member's social lives, both in the context of members of the Pittsburgh community and as members of the Tuscan Protection League, the last of which can be explored more extensively in tandem with the social network analyses.